<p>Gridded precipitation products are inherently uncertain and predominantly deterministic, which limits their applicability in data assimilation systems and hydrologic modeling. This limitation is significant in developing countries such as India, where the observation network is sparse and non-uniform, topography is complex, and hydrometeorological extremes are frequent. The current official 0.25° observed precipitation dataset of the Indian Meteorological Department (IMD) is deterministic and based on Shephard’s interpolation technique. To address these challenges, we have developed the Indian Precipitation Ensemble Dataset (IPED) leveraging the largest network of precipitation gauge stations across India and using a locally weighted spatial regression approach. IPED is a daily 30-member ensemble precipitation product available at 0.1° and 0.25° resolution (1991–2020), accounting for topographical variation in elevation, slope, and aspect. For all thresholds, including the extreme 99^th percentile precipitation during monsoon, the developed ensemble product exhibits higher discrimination and reliability. This is the first observation-based ensemble precipitation product over India and is expected to have widespread hydrometeorological applications.</p>